Elsevier

Atmospheric Environment

Volume 41, Issue 30, September 2007, Pages 6406-6420
Atmospheric Environment

Experimental and numerical study of wind flow behind windbreaks

https://doi.org/10.1016/j.atmosenv.2007.01.014Get rights and content

Abstract

The shelter effect of a windbreak protects aggregate piles and provides a reduction of particle emissions in harbours. RANS (Reynolds-averaged Navier–Stokes equations) simulations using three variants of kε (standard k–ε, RNG k–ε and realizable k–ε) turbulence closure models have been performed to analyse wind flow characteristics behind an isolated fence located on a flat surface without roughness elements. The performance of the three turbulence models has been assessed by wind tunnel experiments. Cases of fences with different porosities (φ) have been evaluated using wind tunnel experiments as well as numerical simulations. The aim is to determine an optimum porosity for sheltering effect of an isolated windbreak. A value of 0.35 was found as the optimum value among the studied porosities (φ=0, 0.1, 0.24, 0.35, 0.4, 0.5).

Introduction

Large amounts of materials in aggregate form are handled, stored and transported in harbour areas producing important particle emissions to the environment. These materials existing in harbours are usually stored piled up on ground in an open location. The open storage produces dust emission from aggregate piles, wind erosion being the main factor of the emissions. These emissions should be controlled and reduced as much as possible and a way to diminish them is to use windbreaks to protect the piles from wind erosion (Borges and Viegas, 1988; Jensen et al., 2001).

The windbreak objective is to provide a shelter effect by decreasing the wind speed in a large zone behind the fence. Depending on barrier characteristics, different wind speed reductions and turbulence features are found in different leeward areas. Porosity, defined as the ratio between the area of the holes and the total area of the fence, is considered as the most influential parameter in the flow pattern behind a windbreak (Perera, 1981). A no re-circulating flow zone behind the windbreak is found when the porosity is higher than 30% (Perera, 1981). Several empirical relationships link the value of the fence resistance coefficient (defined in Section 2.2) with the windbreak porosity (Reynolds, 1969; Perry et al., 1997; Lee and Lim, 2001). The ratio hb/z0, where hb is the height of the barrier and z0 is the roughness length around the windbreak, is also important, especially for modelling purposes (Raine and Stevenson, 1977; Judd et al., 1996).

The literature is rich with studies on windbreak flow with both experiments and numerical simulations. For example, windbreak aerodynamics was reviewed by Plate (1971) and measurements of mean velocity and turbulence variables behind single solid and porous fences were carried out by Raine and Stevenson (1977). The authors found a protection zone for medium porosity fences larger than that for solid windbreaks. In addition, many numerical studies on flow with shelterbelts have been carried out. An early work was made by Tani (1958) where windbreak was treated as a source of momentum deficit, which was considered as a scalar controlled by turbulent diffusion. More complex models have been used with the advent of growing computational power. Wilson (1985) used Reynolds-averaged Navier–Stokes (RANS) equation introducing a momentum sink involving the fence resistance coefficient to simulate a porous barrier. His results showed a quite good prediction of the flow pattern near a single porous windbreak, but further downwind underestimated the rate of return towards the upstream equilibrium. Similar simulations are carried out by Wang and Takle (1995) and Wilson and Mooney (1997). Wang and Takle (1995) did not find the previous problem outlined by Wilson (1985), and Wilson and Mooney (1997) proposed that the reason may be the computational domain size used (too shallow for Wang and Takle, 1995). RANS model inaccuracy in this context has been analysed by Wilson and Yee (2003) and Wilson (2004b). Higher inaccuracy were found in complex cases, as the wind over a windbreak array (data of McAneney and Judd, 1991 case) or the oblique and stratified winds around a shelterbelt (data of Wilson, 2004a case). They found that the RANS models produce an ambiguity in their results due to the choice of the turbulent closure and the discretisation error. They concluded that a balanced view of the problem is the reassessment of both testing closure models and schemes in order to determine the potential value of RANS models and to avoid the premature substitution of measurements for models. For this reason, one part of this study is devoted to perform the comparison of three different turbulence models with the same numerical procedure and parameters (same domain, grid, boundary conditions, etc.) focusing on the differences due to the performance of the turbulence and leaving aside the effects of other factors (grid, boundary conditions, etc.). Other more complex models such as large-eddy simulations (LES) have also been applied. For example, Patton et al. (1998) used LES to simulate turbulent flow field around multiple windbreaks within a homogeneous plant canopy.

This paper is focused on the shelter effect of windbreaks located in harbour areas. In these areas, the fences are used for protecting aggregate piles stored in open locations, and therefore the wind profiles impinging the fences are characterised by roughness lengths corresponding to sea or flat coastal zones which have very low values (10−3–10−4 m) as pointed out by Dyrbye and Hansen (1997) and Gao et al. (2000). To obtain such flat profile the wind tunnel experiments have been carried out without roughness elements obtaining streamwise velocity profiles which corresponds to a z0 close to 0 (see Section 4.1) (even if roughness length can never reach 0, obviously). The shelter effect of an isolated windbreak depends on its porosity, and this paper aims to determine its optimum value giving rise to the largest shelter zone in the above mentioned conditions. Thus, wind flow around isolated windbreaks is studied by means of wind tunnel experiments and numerical simulations for a range of porosities from 0 (solid fence) to 0.5. The experiments have been carried out in the A9 wind tunnel of IDR/UPM, E.T.S.I. Aeronáuticos, Universidad Politécnica de Madrid and the numerical simulations have been performed by a RANS model using different variant of k–ε turbulence closure: standard k–ε, RNG k–ε and realizable k–ε. Thus, each turbulence model results have been compared against wind tunnel data and their influence on the final results has been analysed in the turbulence models. All the other characteristics of the numerical method (grid, domain, etc.) have been maintained identical.

Section snippets

Numerical methods

FLUENT CFD software was used to simulate wind flow patterns around windbreaks.

Bradley and Mulhearn's fence

In past simulation exercises, several studies (Lee and Lim, 2001; Wilson and Yee, 2003) have used the experimental data of Bradley and Mulhearn (1983), denoted B&M data hereafter. Thus, in order to show the possible applicability of the turbulence models used later (standard k–ε, RNG k–ε and realizable k–ε) for solving these cases, we have chosen also B&M data as a benchmark. Here, we only compared experimental and modelled mean flow velocity, while in the IDR/UPM case, studied later in Section

Shelter effect in fences with different porosity

The main objectives of this study are: (1) to find the optimum value of porosity in these conditions based on peak velocity ratio results (defined in Section 4.2), and (2) to evaluate different turbulent closure model performances (standard k–ε, RNG- k–ε and realizable k–ε) using the otherwise identical numerical methods (same grid, domain, boundary conditions, etc.).

Conclusions

The flow around a windbreak located on a flat surface (roughness length almost 0) has been investigated by numerical and experimental studies. The shelter effect produced by an isolated porous fence has been analysed in terms of the peak velocity parameter (Ru1(x,z)), focusing on the fence porosity that produces the largest protection region.

The differences in terms of mean velocity among the three turbulent closure models and B&M results considered are not very large. Larger differences are

Acknowledgements

The authors wish to thank CIEMAT (Spain) for doctoral fellowship held by J.L. Santiago, and to Puertos del Estado (Spain) for their support concerning the wind tunnel experiments. We also thank A. Martilli for important comments on the manuscript.

References (40)

  • M.D.A.E.S. Perera

    Shelter behind two-dimensional solid and porous fences

    Journal of Wind Engineering and Industrial Aerodynamics

    (1981)
  • E.J. Plate

    The aerodynamics of shelter belts

    Agricultural Meteorology

    (1971)
  • P.J. Richards et al.

    Appropriate boundary conditions for computational wind engineering models using the k-ε turbulence model

    Journal of Wind Engineering and Industrial Aerodynamics

    (1993)
  • T.H. Shih et al.

    A new k–ε eddy-viscosity model for high Reynolds numbers turbulent flows—model development and validation

    Computers Fluids

    (1995)
  • J.D. Wilson

    Numerical studies of flow through a windbreak

    Journal of Wind Engineering and Industrial Aerodynamics

    (1985)
  • J.D. Wilson et al.

    Calculation of wind disturbed by an array of fences

    Agricultural and Forest Meteorology

    (2003)
  • J.B. Barlow et al.

    Low-speed Wind Tunnel Testing

    (1999)
  • B.F. Bradley et al.

    Development of velocity and shear stress distributions in the wake of porous shelter fence

    Journal of Wind Engineering and Industrial Aerodynamics

    (1983)
  • C. Dyrbye et al.

    Wind Effect on Structures

    (1997)
  • C. Frank et al.

    Double-arranged mound-mounted shelterbelts: influence of porosity on wind reduction between the shelters

    Environmental Fluid Mechanics

    (2005)
  • Cited by (127)

    View all citing articles on Scopus
    View full text